Read doi:10.1016/j.enggeo.2007.01.008 text version

Engineering Geology 91 (2007) 178 ­ 193 www.elsevier.com/locate/enggeo

Microtremor HVSR study for assessing site effects in the Bovec basin (NW Slovenia) related to 1998 Mw5.6 and 2004 Mw5.2 earthquakes

Andrej Gosar

Environmental Agency of Slovenia, Seismology and Geology Office, Dunajska 47, SI-1000 Ljubljana, Slovenia University of Ljubljana, Faculty of Natural Sciences and Engineering, Slovenia Received 28 July 2006; received in revised form 9 January 2007; accepted 25 January 2007 Available online 1 February 2007

Abstract The Bovec basin, which is filled with glacial and fluvial sediments, has recently been struck by two strong earthquakes (1998 and 2004) which caused extensive damage (VII­VIII EMS-98). Strong site effects resulted in large variations in damage to buildings in the area, which could not be explained by the surface variations in Quaternary sediments. The microtremor horizontal-to-verticalspectral ratio (HVRS) method was therefore applied to a 200 m dense grid of free-field measurements to assess the fundamental frequency of the sediments. Large variations in the sediment frequency (3­22 Hz) were obtained, with most of the observed values in the range 6­12 Hz. The observed frequencies cannot be related to the total thickness of Quaternary sediments (sand, gravel), but can be explained by the presence of conglomerate or lithified moraine at shallow depths. The results were compared also with the velocity structure derived from seismic refraction data. Microtremor measurements performed in several two and some three- and four-storey houses (masonry with RC floors), which prevail in the Bovec basin, have shown that the main building frequencies in the area are in the range 7­11 Hz. This indicates that damage to houses in both earthquakes in some parts of the basin was enhanced by site amplification and soil-structure resonance. Areas of possible soil-structure resonance were identified in the settlements Bovec­ Brdo, Bovec­Mala vas, Cezsoca and Kal-Koritnica. Considerable changes in fundamental frequencies within short distances were established in the town of Bovec. Their values are as high as 22 Hz in the central part of the town, but diminish to 6­11 Hz in the adjacent Brdo and Mala vas districts. This is in agreement with the distribution of damage in both earthquakes, which was considerably higher in Brdo and Mala vas, although the houses in the central part of the town are older. Microtremor investigations have proved an effective tool for assessment of site effects in cases of complex geological structure commonly encountered in young Alpine basins filled with glaciofluvial sediments which are partly cemented. Lithified layers can considerably change the fundamental frequency and, consequently, the site effects. By taking additional measurements in buildings possible soil-structure resonance can be identified. © 2007 Elsevier B.V. All rights reserved.

Keywords: Ambient vibrations; Microtremors; Horizontal-to-vertical spectral ratio (HVSR); Site effects; Soil-structure resonance

1. Introduction The Bovec basin is located in the Upper Soca valley in NW Slovenia (Fig. 1), a region undergoing a recent increase in seismic activity. Two strong earthquakes

E-mail address: [email protected] 0013-7952/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.enggeo.2007.01.008

A. Gosar / Engineering Geology 91 (2007) 178­193

179

struck the area in 1998 and 2004. Both earthquakes occurred on the NW­SE trending near-vertical Ravne fault in the Krn mountains at 7­9 km depth and caused extensive damage to buildings in the area. The focal mechanisms of both earthquakes show almost pure dextral strike­slip. The epicentral distance to the town of Bovec was 6­7 km. The 12 April 1998 earthquake (Mw = 5.6) had a maximum intensity of VII­VIII EMS98 (Zupancic et al., 2001) and the 12 July 2004 (Mw = 5.2) earthquake VI­VII EMS-98 (Zivci et al., 2006). Strong variations in damage to buildings were observed within short distances in the whole Bovec basin. They cannot be explained by changes in the

epicentral distance or by changes in the radiation of seismic energy from the source, although the latter can have some influence and needs further attention. The variations in damage can only be attributed in part to differences in building vulnerability, since the building typology is similar throughout the area. Local geological conditions (site effects) therefore played the most important role. NW Slovenia marks a kinematic transition between E­W striking thrust faults of the Alpine system and NW­SE striking faults of the Dinarides system. Before 1998, relatively weak rates of seismicity were characteristic of this area. Nevertheless, in the seismic

Fig.1. Shaded relief map of the Upper Soca valley with the Bovec basin and epicentres of two recent strong earthquakes in Krn mountains. Fault plane solution for 12/04/1998 event is from Zupancic et al. (2001) and for 12/07/2004 event from Kastelic et al. (2006). Boxed area indicates the map shown in Fig. 2.

180

A. Gosar / Engineering Geology 91 (2007) 178­193

hazard map for a return period of 475 years, the Bovec basin is located in a region with increased seismic hazard, with 0.225­0.250 g design ground acceleration (Lapajne et al., 2001). This is mainly due to the proximity of high level of seismic activity in the Friuli area (30­40 km to the W) where the Mw = 6.4 Friuli earthquake occurred in 1976 (Perniola et al., 2004). The strongest earthquake ever recorded in the Alps­ Dinarides junction was the 1511 western Slovenia earthquake (M = 6.8); the exact location and mechanism of this event are still debated (Fitzko et al., 2005). Seismic microzonation study of the Bovec basin based on surface geological data and data from shallow geotechnical boreholes (Ribicic et al., 2000) has shown that it is not possible to explain with these data most of the observed variations in the distribution of damage. More promising results have been obtained by combined application of microtremors and 1D modelling of ground motion based on the results of shallow

geophysical investigations (Gosar et al., 2001). Some locations of probable soil-structure resonance were identified in this study and large variations in damage to buildings related to the 1998 earthquake were explained by significant variations in ground motion amplification between two different parts of the town of Bovec. To study in more detail the effects of surface geology on seismic ground motion, we performed free-field microtremor measurements at 124 points in the Bovec basin. In addition, 20 buildings were surveyed with the microtremor method to determine the building main frequencies. Horizontal-to-Vertical Spectral Ratio (HVSR) analysis was performed to determine the fundamental frequency of the sediments. The main building frequencies were identified from microtremor measurements performed on different floors inside the buildings. By overlying the frequency map of the sediments with the frequencies of buildings, an attempt

Fig. 2. Simplified geological map of the Bovec basin after Buser (1986), Jurkovsek (1986) and Bavec et al. (2004). Transfer to GIS was done by Geological Survey of Slovenia. Black polygons are settlements. Boxed area indicates the map shown in Fig. 4.

A. Gosar / Engineering Geology 91 (2007) 178­193

181

was made to identify the areas of potential soil-structure resonance, which was in our opinion the main reason for the relatively high damage to buildings in different parts of the Bovec basin. Results of microtremor investigations were compared to the distribution of damage caused by the 1998 and 2004 earthquakes and to available geological, geotechnical and geophysical data. 2. Geological setting The geology of the Bovec basin and its surroundings is very complex and still not resolved in detail as far as the structural geology is concerned. The Bovec basin (6 km long and 2 km wide) developed in the Alpine valley of the Soca river, which is elsewhere very narrow (Fig. 1). The basement consists of Mesozoic platform carbonates of Upper Triassic and Jurassic age (Fig. 2). They are overlain by a succession of deep-water flysch or by marly limestone (scaglia) with intercalated calcarenites, shales, marls and conglomerates of Cretaceous age (Buser, 1986; Jurkovsek, 1986). Quaternary sediments are represented from bottom to top by partly lithified glaciofluvial sediments, overlain by lacustrine chalk (Bavec et al., 2004). During the Holocene, the chalk was partly eroded (in some areas totally) and covered by glaciofluvial sand and gravel which are weakly cemented in some parts into conglomerate and by unconsolidated moraine (till). The lithologic column can be divided into three sections, based on physical properties. The basement consists of carbonates. Upper Triassic dolomite and limestone are overlain by Liassic limestone (Fig. 2). Normal erosional contacts with a succession of deep water clastics (flysch) of Jurassic and Cretaceous age are found all around the steep flanks of the Bovec basin. In the eastern part, the succession of deep water sediments

starts with carbonate breccia ("wild" flysch), whereas at other locations platform carbonates are overlain by marly limestone (scaglia) and intercolated calcarenites, shales, marls and conglomerates (Buser, 1986; Jurkovsek, 1986). Quaternary sediments represent the third section. A thick sequence of lacustrine chalk covers lithified glacial and fluvial sediments in the central part of the basin (W of the map in Fig. 2); (Kuscer et al., 1974; Bavec, 2002). Glacial moraine sediments are in parts cemented (tillite). During the Holocene, the chalk was partly eroded and covered by older (Bovec terrace) and younger sand and gravel, which are in some parts weakly cemented in conglomerate, by scree and debris under the steep flanks of the basin and by recent fluvial sediments along the Soca river (Bavec, 2002). A 2D cross-section (Fig. 3) was prepared based on general geological knowledge of the area and sparse geophysical data, because no deep boreholes have been drilled in the basin. Electrical sounding was used to determine the thickness of glaciofluvial sediments, but it did not distinguish between flysch and lacustrine chalk (Gosar et al., 2001). In the vicinity of the cross-section, according to the geological data, the chalk is found mainly in the Cezsoca area under younger sand and gravel, while under the Bovec terrace only flysch is expected. Seismic velocities of different lithological units were derived from shallow seismic refraction measurements using P and S-waves and from down-hole seismic velocity measurements in geotechnical boreholes. Both were performed only in the border part of the basin, where the main settlements are located, as part of geotechnical studies for retrofitting damaged houses. The S-waves velocity for glaciofluvial deposits are 250­ 450 m/s, for till 300­600 m/s, for tillite 600­800 m/s and for flysch 1000­1400 m/s (Gosar et al., 2001).

Fig. 3. Cross-section of Quaternary sediments across the Bovec basin interpreted from vertical electrical soundings and microtremor HVRS curve derived from the map in Fig. 4 where the position of the cross-section is also shown.

182

A. Gosar / Engineering Geology 91 (2007) 178­193

3. Microtremor measurements and analyses The microtremor HVSR method is widely used for microzonation and site effect studies. Reviews on the method can be found in Bard (1999) or Mucciarelli and Gallipoli (2001). However, the theoretical basis of HVSR method is still debated and different explanations have been given. Nakamura's (2000) "body waves" explanation is based on S-wave resonance in soft sediment layer with minor or neglecting influence of surface waves. More widely accepted (Bard, 1999; Bonnefoy-Claudet et al., 2006) is "surface waves" explanation that HVSR is related to the ellipticity of Rayleigh waves which is frequency dependent. HVSR exhibits therefore a sharp peak at fundamental frequency of the sediments, if there is a high enough impedance contrast between sediments and bedrock. Criticism of the method was often related to the fact that there is no common practice for data acquisition and processing (Mucciarelli and Gallipoli, 2001). Attempts to provide standards were only recently been made (SESAME, 2004). It is widely accepted today that the frequency of the HVSR peak reflects the fundamental frequency of the sediments. Its amplitude depends mainly on the impedance contrast with the bedrock and cannot be used as a site amplification. However, comparison with results of standard spectral ratio method has shown that HVSR peak amplitude underestimates the actual site amplification (Bard, 1999; SESAME, 2004). HVSR also does not provide any estimate of the actual bandwidth over which the ground motion is amplified. The main advantages of HVSR method are simple and low cost measurements which can be performed at any time and any place, and direct estimate of sediments resonance frequency without knowing geological and S-velocity structure of the underground. Any knowledge of thickness or/and velocity of sediments and comparison with other methods and earthquake damage can significantly improve the reliability of the results (Bard, 1999). Since most noise modelling simulations were done for 1D cases (BonnefoyClaudet et al., 2006) the influence of 2D and 3D structures on HVSR is not yet adequately explored. On the other hand, by using microtremors much denser grid of measurements is possible than with any method based on earthquake recording, on geophysical investigations or on drilling. The use of microtremors was later extended to identifying the main frequencies of the buildings, their vulnerability and soil-structure resonance (Mucciarelli et al., 2001; Gallipoli et al., 2004a). Damage enhancement and soil-structure resonance was recently studied using microtremors for Umbria­Marche earthquake (Mucciar-

elli and Monachesi, 1998; Natale and Nunziata, 2004), for Thessaloniki earthquake (Panou et al., 2004) and for Molise earthquake (Gallipoli et al., 2004b). In the Bovec basin, measurements of ambient vibrations were performed by using six portable seismographs Tromino (Micromed) composed of three orthogonal electrodynamic velocity sensors, GPS receiver, digitizer and recording unit with flash memory card. All parts are integrated in a common case to avoid electronic and mechanical noise that can be introduced by wiring between equipment parts. 3.1. Free-field Good ground coupling on soft soils was obtained by using long spikes mounted at the base of the seismograph. The recording length was 20 min, which allows spectral analysis down to 0.5 Hz. In the Bovec basin, a 7 km 2 large area was surveyed in an approximate 200 × 200 m grid (Fig. 4). This area covers almost the whole extent of Quaternary sediments in the central part of the basin and includes Bovec, Cezsoca and Kal-Koritnica (Figs. 2 and 4), the three settlements which suffered extensive damage in the 1998 and 2004 earthquakes. Altogether, 124 free-field measurements were performed. Their locations were carefully selected to avoid the influence of trees, sources of monochromatic noise, rivers and strong topographic features (edges of terraces). Measurements were performed only on no-wind days, because it is known, and confirmed also with our tests, that the noise introduced by strong wind severely affects the reliability of HVSR analysis (SESAME, 2004; Mucciarelli et al., 2005). The noisiest conditions were in the Cezsoca area, because of retrofitting construction activities following the 2004 earthquake. Coherent noise in the vicinity of the industrial area in the easternmost part of Bovec was found to be less problematic, because the dominant frequency of this noise was seldom below 20 Hz. HVSR analysis was performed in the following way. Recorded time series were visually inspected to identify possible erroneous measurements and stronger transient noise. Each record was then split into 30 s long nonoverlapping windows for which amplitude spectra in a range 0­64 Hz were computed using a triangular window with 5% smoothing and corrected for sensor transfer function. HVSR was computed as the average of both horizontal component spectra divided by the vertical spectrum for each window. From the colour coded plot of HVSR functions for all 40 windows, the windows including strong transient noise were identified in order to be excluded from further computation.

A. Gosar / Engineering Geology 91 (2007) 178­193

183

Fig. 4. Contour map of the fundamental frequency peak derived from HVSR analysis of microtremor data. Frequency range of possible soil-structure resonance (6­12 Hz) is shaded. Triangles indicate points of free-filed measurements; labelled triangles are examples of HVSR analyses shown in Figs. 6 and 7. H1 to H6 indicate measurements in houses shown in Figs. 9­14. A­A indicates cross-section shown in Fig. 3.

The effect of transient seismic noise on HVSR analysis is still debated. Parolai and Galiana-Merino (2006) have shown that transients have no or very little effect on the HVSR. In our measurements transient noise occurs mostly in a frequency range lower than is the value of the peak. At the end, the average HVSR function with a 95% confidence interval was computed. In addition, a directional HVSR analysis was performed in 10° angular steps to identify possible preferential directions of noise sources. 3.2. Measurements inside buildings For measurements inside buildings, shorter spikes mounted at the bottom of the seismograph were used to enable precise levelling, but to avoid vibrations of the unit. The recording length was 10 min, because frequencies below 1 Hz were not of interest. Two-storey residential houses prevail in the area, with some multiflat 3- or 4-storey buildings. Older houses are mainly of simple and massive stone, while newer houses are

mainly masonry with RC floors. Measurements were performed on all floors of the building. Two horizontal components were oriented in the longitudinal ("N­S" component of the seismograph) and transverse ("E­W" component) directions of the building. The instrument was placed as close as possible to the mass centre of the building and close to the wall. Close to each building, but far enough to avoid its influence, a free-field measurement was also performed. Each record was split into 10 s long non-overlapping windows for which amplitude spectra were computed using a triangular window with 3% smoothing. Windows including strong transient noise were excluded from further computation, although some investigations indicate that influence of transients is small (Yuen et al., 2002). Average amplitude spectra for each component were computed from selected windows. For each floor, a separate HVSR of each horizontal component to the vertical component was also computed, as proposed by Gallipoli et al. (2004a). Directional analysis was performed in 10° angular steps for easier recognition of

184

A. Gosar / Engineering Geology 91 (2007) 178­193

the two main frequencies in longitudinal and transverse directions. In addition, ratios of amplitude spectra for records taken on higher floors to the spectra of a reference record taken at ground floor were computed. 4. Results and interpretation 4.1. Free-field results and the map of fundamental frequency of sediments HVSR analyses of 124 free-field microtremor measurements in the Bovec basin showed that most of them (80%) fulfill the criteria defined by SESAME project (Table 1) for reliable measurements (SESAME, 2004). Three criteria for a reliable HVSR curve are based on the relation of a peak frequency to the window length, number of significant cycles and standard deviation of a peak amplitude. Six criteria for a clear peak are based on the relation of the peak amplitude to the level of the HVSR curve elsewhere, and standard deviations of the peak frequency and of its amplitude (the amplitude should decrease rapidly on each side). If all three criteria for reliable curve and at least five criteria for a clear peak are fulfilled, the frequency of the peak is considered to be the fundamental frequency of sediments down to the first stronger contrast in shear-wave velocity. The main reasons for the other 20% of measurements were: a) artificial noise, b) two or more peaks in a spectrum or c)

Fig. 5. Amplitude vs. frequency graph of HVSR peaks.

too small amplitude of the peak. In cases in which the small amplitude of the HVSR peak caused failure to the criteria for a clear peak, we compared the results with adjacent measurements. If the frequencies of questionable peaks were comparable with the frequencies obtained at adjacent points, we kept them in the database. Finally, the frequencies of peaks determined at 104 points were used for contouring a map of the fundamental frequency of sediments in the Bovec basin (Fig. 4). The frequencies of the observed HVSR peaks are distributed in the wide range of 3­22 Hz (Fig. 5), but

Table 1 Criteria for reliable HVSR curve and clear HVSR peak defined by SESAME project (SESAME, 2004)

A. Gosar / Engineering Geology 91 (2007) 178­193

185

most of them (46%) are in the range 7­11 Hz, 14% are below 5 Hz and 6% above 20 Hz. Since two-storey residential houses prevail in the area, their vulnerability can be expected in the frequency range 6­12 Hz as is described in the next section. The area with fundamental frequency of sediments in this range is shaded on the map (Fig. 4). The danger of soil-structure resonance in the Bovec basin therefore seems to be considerable. This is discussed further in the section describing the results of measurements in buildings. The amplitudes of the HVRS peaks are distributed in a range of 2­12, but most of them (80%) are in the range 3­6, 10% in the range 6­8, 5% below 3 and 5% above 10. No correlation between the frequency of the peak and its amplitude was established (Fig. 5). Comparison of the HVSR frequency map (Fig. 4) with the geological map (Fig. 2) and geological crosssection (Fig. 3) showed that large variations in fundamental frequencies cannot be explained by known variations in surface geology or related to the total thickness of the Quaternary sediments. Seismic refraction measurements with P and S-waves performed in the frame of geotechnical investigations for retrofitting at three locations in urban areas of the Bovec basin (Table 2) allowed us to compare the fundamental frequency obtained from microtremors with the calculated values for sediments for assumed single layer. At all three points is the S-velocity in the bedrock above 800 m/s and the impedance contrast with average velocity in sediments (300­390 m/s) is between 2 and 3. There is a good correspondence between microtremors HVSR peak frequency and calculated values. This allowed us to make some interpretations of fundamental frequencies obtained from microtremors in terms of shallow geological structures for expected lithological units. It is clear that the microtremor fundamental frequencies do not reflect in the central part of the basin the depth to flysch or deep sea clastics which is up to 100 m. If a mid S-wave velocity for glaciofluvial sediments of 350 m/s is taken for sediments, the rough estimate of the corresponding

thickness of assumed single layer for observed frequencies (3­22 Hz) would give a range of 4­30 m. Selected results of HVRS analyses measured close to the cross-section A­A (Fig. 3) are shown from N to S in Fig. 6. For the E part of Bovec­Mala vas (measurements B33, B37a, B44) frequencies around 8 Hz are characteristic, with relatively strong amplitude. For instance, at point B33, a very clear peak with an amplitude of 12 was obtained. Towards the W (centre of Bovec), there is a steep increase of frequency. For instance, at B41 a peak at 20.8 Hz with amplitude of 4.5 was obtained. In the central part of the basin (B47 and B92), frequencies around 10.5 Hz were obtained, with wider peaks and amplitudes between 4 and 5. Monochromatic noise (22 Hz) is clearly seen at B47. B105 is an example of a measurement with a low amplitude and a wide peak. Points B186 and B99 have some common characteristics, although the first was measured at 80 m higher elevation on the Bovec terrace than the second one (close to the Soca river). In both HVSRs, several smaller peaks are present inside the wide range of amplification. This range is between 4 and 6 Hz at B186 and between 12 and 20 Hz at B99. In Cezsoca (B159a), noise due to construction activities severely affected the measurements. Spectral ratios were therefore in general of lower quality in this area. The curve of fundamental frequencies derived from the map (Fig. 4) is shown at the top of the cross-section A­A (Fig. 3). Its main characteristics are: · Frequencies of 8 to 10 Hz in the southern part (Cezsoca) correspond to a 15­25 m thick layer of younger sand and gravel overlaying lacustrine chalk and flysch. · The 12.5 Hz peak close to the Soca river is most probably related to shallow lacustrine chalk. · Frequencies of 10 to 14 Hz observed in the northern part of the profile are probably related to the layer of conglomerate inside the Bovec terrace. The extent of this weakly cemented layer was not known before, because it was not detected by electrical sounding

Table 2 Seismic refraction data on thickness and S-velocity of sediments and comparison of microtremor HVSR peak frequency with calculated frequency for single layer Location Microtremor HVSR f0 Thickness (h) B37a (Mala vas) H2 (Brdo) B99 (Cezsoca) 8.0 Hz 9.4 Hz 13.3 Hz 9.0 ± 1.0 m 9.9 ± 1.4 m 6.9 ± 0.4 m Seismic refraction S-velocity (Vs) 300 m/s 390 m/s 350 m/s f0 = Vs/4h 8.3 ± 0.8 Hz 9.8 ± 1.4 Hz 12.7 ± 0.7 Hz

186

A. Gosar / Engineering Geology 91 (2007) 178­193

Fig. 6. Selected microtremor HVSR analyses in the vicinity of the cross-section shown in Fig. 3. Thin lines representing 95% confidence interval are shown.

due to the small resistivity contrast. The only known exposures of conglomerate are in the cliff of the Bovec terrace on the N bank of Soca river (Fig. 8). · The largest variations in frequency (6­14 Hz) on this profile were observed in the Bovec area. The town is located at the margin of the Bovec terrace, partly on it

(sand, gravel), partly on glacial till and tillite overlying flysch rocks. In any case, the thickness of Quaternary sediments is in general small, but due to fluvial relief the variations can be considerable. Shallow seismic refraction investigations have shown that shear-wave velocity can also vary considerably, depending on

A. Gosar / Engineering Geology 91 (2007) 178­193

187

whether the sediments are cemented (tillite or conglomerate). This section describes some examples of HVSR analyses shown in Fig. 7 and the main features of the map of HVSR fundamental frequencies of the Bovec basin (Fig. 4).

· Frequencies in the range 7­11 Hz prevail (B33, B37a, B44 and B92 in Fig. 6) in the largest part of the map (Bovec terrace, Cezsoca, Kal-Koritnica). Although some peaks are quite wide (B118 in Fig. 7) or the measurements disturbed by monochromatic noise (B145 in Fig. 7), it was possible to determine the

Fig. 7. Some examples of microtremor HVSR curves from the Bovec basin. Locations of measurements are shown in Fig. 4. Thin lines representing 95% confidence interval are shown.

188

A. Gosar / Engineering Geology 91 (2007) 178­193

Fig. 8. The cliff of the Bovec terrace (N bank of Soca river) near point B117 which is located 100 m from cliff edge. Weakly cemented conglomerate is visible near top of cliff.

fundamental frequency in the majority of spectral ratios. · Lower frequencies (3­6 Hz) were obtained at the margins of the basin (W of Bovec (B22), E of Cezsoca (B171), W of Kal-Koritnica (B56)) where flysch and deep sea clastic rocks outcrop. An isolated point (B186 in Fig. 6) with low frequency of 4.2 Hz was obtained also south of the airport, but the HVSR peak is not very reliable. · Very high frequencies (18­22 Hz) were obtained in three areas. 1) In the central part of Bovec they are related to a very thin layer of till or sand­gravel overlying flysch bedrock. 2) At the SE margin of the map (B117, B124), they are related to very shallow conglomerate visible in a nearby cliff (Fig. 8). 3) At point B106, the peak is wide and has a low amplitude, but is most probably also related to a shallow layer of conglomerate in the Bovec terrace. Comparison of several adjacent measurements has shown that the amplitudes of the HVSR peaks are also highly variable within short distances. For example, points B29 and B81 (Fig. 7), which were measured only 200 m away, show almost the same fundamental frequency (6 and 6.2 Hz), but have considerably different peak amplitudes (4.5 and 11). 4.2. Measurements in buildings and soil-structure resonance Measurements performed in buildings confirmed that microtremors are an effective tool for identification of the main building frequencies. For all measured houses

(Figs. 9­14), it was possible to identify the longitudinal and transverse frequencies. In buildings "N­S" component corresponds to its longitudinal and "E­W" component to its transverse direction. We used amplitude

Fig. 9. Example of possible soil-structure resonance for two-storey house H1 in the Bovec town­Mala vas area (Fig. 4). a) free-field HVSR, b) 2nd floor of the house -- HVSR of each horizontal component to the vertical component c) 2nd floor of the house -- amplitude spectra of horizontal components.

A. Gosar / Engineering Geology 91 (2007) 178­193

189

components. The results of free-field measurements performed in the vicinity of buildings were always carefully inspected and compared to adjacent measurements to avoid possible erroneous results due to the influence of nearby structures. For each house, damage caused by the 1998 and 2004 earthquakes, assessed according to the classification of damage for masonry buildings in European Macroseismic Scale (ESM-98), is also given. House H1 (Fig. 9) is located in Bovec­Mala vas district on the Bovec terrace, where a clear peak at frequency 8 Hz with amplitude of 6.5 was obtained in free-field measurements. The old house at this location was very heavily damaged in the 1998 earthquake. It was therefore demolished and a new house of similar size was built at the same location. In the 2004 earthquake, the new house suffered slight to moderate damage. From microtremor measurements in the house, it was possible to identify the main building longitudinal (8.7 Hz) and transverse (7.6 Hz) frequencies. Soilstructure resonance is therefore a probable reason for the very high damage to the previous house in the 1998

Fig. 10. Example of possible soil-structure resonance for two-storey house H2 in the Bovec town­Brdo area (Fig. 4). a) free-field HVSR, b) 2nd floor of the house -- HVSR of each horizontal component to the vertical component c) 2nd floor of the house -- amplitude spectra of horizontal components.

spectra of the horizontal components or HVSRs of each horizontal component to the vertical component as proposed by Gallipoli et al. (2004a). In general, by using the spectral ratio, narrower and more prominent peaks were obtained than by using simple amplitude spectra (see Fig. 12 for example), but the peaks are already clear enough in the amplitude spectra. There was good correspondence in the frequency of the peaks between measurements taken on different floors in the building, showing an increasing response at higher levels. Since it is not the purpose of this paper to discuss the dynamic behaviour of buildings, but to identify possible soil-structure resonance, only the results from the highest floor of each building are presented below. Since two-storey residential houses (masonry with RC floors) prevail in the area, examples of five twostorey and one three-storey buildings from different parts of the Bovec basin are presented in this section. For each building the following graphs are presented: a) the results of HVSR analysis for free-field measurement performed in the vicinity of the building, b) HVSR of each horizontal component to the vertical component for the highest floor, c) amplitude spectra of both horizontal

Fig. 11. Example of two-storey house H3 in the Bovec town­Brdo area (Fig. 4) where soil-structure resonance is not likely. a) free-field HVSR, b) 2nd floor of the house -- HVSR of each horizontal component to the vertical component c) 2nd floor of the house -- amplitude spectra of horizontal components.

190

A. Gosar / Engineering Geology 91 (2007) 178­193

fact, the house was only slightly damaged in both earthquakes. House H4 (Fig. 12) is located in Cezsoca on younger Quaternary gravel, where a wide peak (6.5­ 10 Hz) centred around 8 Hz was obtained in free-field, which were partly disturbed by monochromatic noise (4­4.5 Hz). The main building longitudinal (10.2 Hz) and transverse (8.5 Hz) frequencies were identified. Soil-structure resonance is therefore very likely. The house was moderately damaged in the 1998 earthquake, later retrofitted, and then substantially damaged in the 2004 earthquake. In general, the damage in Cezsoca was slightly higher in the 2004 event than in 1988 event. The reason for this can be sought in slightly different mechanism of the Ravne fault rupture from the 2004 event close to the structural barrier of the Bovec basin (Kastelic et al., 2006), but is not yet adequately explained. House H5 (Fig. 13) is located in Kal-Koritnica on the Bovec terrace, where a wide peak centred around 8.6 Hz was obtained in free-field. The main longitudinal (8.3 Hz) and transverse (8.8 Hz) frequencies of the building were identified. The house was moderately

Fig. 12. Example of possible soil-structure resonance for two-storey house H4 in Cezsoca (Fig. 4). a) free-field HVSR, b) 2nd floor of the house -- HVSR of each horizontal component to the vertical component c) 2nd floor of the house -- amplitude spectra of horizontal components.

earthquake, as well as for the relatively high damage in 2004 to the new house constructed according to the latest building codes. In general, the intensity in the 1998 earthquake was higher in Mala vas (VII EMS-98) than in other districts of Bovec (VI EMS-98) (Gosar et al., 2001). House H2 (Fig. 10) is located in Bovec­Brdo district on the Bovec terrace, where a wide peak centred around 9.3 Hz with amplitude of 5.0 was obtained in free-field measurements. The main building longitudinal (8.5 Hz) and transverse (9.7 Hz) frequencies were identified. Soil-structure resonance is therefore very likely. The house was moderately damaged in the 1998 and 2004 earthquakes. House H3 (Fig. 11) is also located in Bovec­ Brdo on the Bovec terrace, 100 m to the north of house H2. A wide peak centred around 17.5 Hz with HVSR amplitude of 5.0 was obtained in free-field measurements. The main building longitudinal (10.7 Hz) and transverse (9.6 Hz) frequencies were identified. Since both frequencies of the building are considerably lower than the frequency of sediments, soil-structure resonance would not be expected. In

Fig. 13. Example of possible soil-structure resonance for two-storey house H5 in Kal-Koritnica (Fig. 4). a) free-field HVSR, b) 2nd floor of the house -- HVSR of each horizontal component to the vertical component c) 2nd floor of the house -- amplitude spectra of horizontal components.

A. Gosar / Engineering Geology 91 (2007) 178­193

191

damaged in the 1998 earthquake, later retrofitted, and again moderately damaged in the 2004 earthquake. Soilstructure resonance is therefore very likely, especially because the damage to other buildings in Kal-Koritnica was considerably smaller during the 2004 event than in 1998. Three-storey multi-flat house H6 (Fig. 14) is located in Bovec­Brdo. A wide peak centred around 11.2 Hz with amplitude of 4.5 was obtained in the free-field and the main building longitudinal (5.8 Hz) and transverse (6.6 Hz) frequency were identified. Since both frequencies of the building are considerably lower than the frequency of sediments, soil-structure resonance would not be expected. In fact, the house was only slightly damaged in both earthquakes. The graph of main building frequency vs. height (no. of storeys) for all 20 examined buildings is shown in Fig. 15. The main building frequency for two-storey houses is considerably higher (7­13 Hz) than for three- and four-storey houses (5­9 Hz). The difference between the later two categories is not significant. The difference in the building main frequency in long-

Fig. 15. Plot of the frequency vs. height (no. of storeys) of 20 examined buildings (11 two-storey, 6 three-storey, 3 four-storey).

itudinal and transverse direction is in the range of 0.5 to 2 Hz. 5. Discussion and conclusions Microtremor investigations have proved to be an effective tool for assessment of site effects in the case of complex geological structures commonly encountered in young Alpine basins filled with glacial and fluvial sediments that are partly cemented. Cemented layers (conglomerate, lithified moraine -- tillite) can considerably change the fundamental frequency. For several reasons: the irregular shape of layers, weak contrast in electrical resistivity, layers are frequently thin, it is very difficult to detect these layers by geophysical methods. For microzonation purposes of larger areas, the cost of geophysical investigations usually also precludes performing measurements in a grid dense enough to reveal the detailed stratigraphy. Drilling is also not always effective in discriminating sand­gravel from weakly cemented conglomerate, because the latter is easily disintegrated by drilling bits. In such geological conditions, the microtremor HVSR method is very useful for quantitative seismic microzonation and assessment of possible soil-structure resonance. At the same time, the costs of measurements and processing are kept low, because no active source is needed. Microtremor measurements performed in the Bovec basin showed large variations in the fundamental frequency (3­22 Hz) of sediments. Because most observed frequencies were in the range 7­11 Hz, they

Fig. 14. Example of three-storey house H6 in the Bovec town­Brdo area (Fig. 4) where soil-structure resonance is not likely. a) free-field HVSR, b) 3rd floor of the house -- HVSR of each horizontal component to the vertical component c) 3rd floor of the house -- amplitude spectra of horizontal components.

192

A. Gosar / Engineering Geology 91 (2007) 178­193

overlap with the range of building frequencies (5­ 13 Hz). This indicated that the danger of soil-structure resonance is considerable in the area. Observed frequencies cannot be related to the total thickness of Quaternary sediments, but can be explained by shallow layers of lithified sand­gravel (conglomerate) or glacial moraine (tillite). The complex geological structure has resulted in some considerable changes in fundamental frequencies within short distances. This was especially pronounced in the town of Bovec, which is located at the margin of the Bovec terrace, partly on sand and gravel, and partly on glacial till and tillite overlaying flysch rocks (Fig. 2). In any case, the thickness of Quaternary sediments is small, but due to fluvial relief the variations can be considerable. The frequencies obtained by microtremor measurements are as high as 22 Hz in the central part of the town and diminish towards Brdo and Mala vas districts, where frequencies in the range 6­11 Hz predominate (Fig. 4). This was reflected in the distribution of damage in both earthquakes, which was lower in the central part of the town although the buildings are older (Gosar et al., 2001). In this study, the first attempt was made to use microtremors for identifying the main frequencies of the buildings in the area and to establish possible soilstructure resonance. The results are promising and indicated that the major variations in the distribution of damage in the Bovec basin can be explained by soilstructure resonance. Measurements in two-storey houses which prevail in the area showed that the main building frequencies in the area are in the range of 7­11 Hz. Areas of possible soil-structure resonance were therefore identified in the settlements of Bovec­Brdo, Bovec­Mala vas, Cezsoca and in Kal-Koritnica. In the whole surveyed area of the Bovec basin, the area with fundamental frequencies of sediments in the range 6­ 12 Hz, in which soil-structure resonance is possible, occupies more than 60% (Fig. 4). Nevertheless, before more general conclusions can be made, microtremor measurements in a larger number of houses should be performed, including analyses of their dynamic behaviour and of available information on the construction of individual buildings. In addition investigations should be enhanced to methods dealing with 2D or even 3D shape of the Bovec basin. Although sparse geophysical data limits the knowledge about the shape of the bedrock, the first attempts of 2D modelling (Vanini et al., 2006) of observed seismic ground motion at single strong motion seismic station located in the Bovec basin at the time of 2004 earthquake gave promising results.

Acknowledgments The study was realised with the support of NATO SfP project 980857: Assessment of seismic site amplification and seismic building vulnerability in FYR of Macedonia, Croatia and Slovenia and Interreg IIIB Alpine Space project SISMOVALP: Seismic hazard and Alpine valley response analysis. The author is indebted to Miha Lubi for his help in field measurements and to Barbara Sket-Motnikar for valuable suggestions. References

Bard, P.Y., 1999. Microtremor measurements: a tool for site effect estimation? In: Irikura, K., Kudo, K., Okada, H., Sasatami, T. (Eds.), The Effects of Surface Geology on Seismic Motion. Balkema, pp. 1251­1279. Bavec, M., 2002. New temporal, and genetic determination of some late Quaternary sediments in Bovec basin and its surroundings (NW Slovenia). Geologija 45, 291­298 (in Slovenian, with English summary). Bavec, M., Tulaczyk, S.M., Mahan, S.M., Stock, G.M., 2004. Late Quaternary glaciation of the Upper Soca River Region (Southern Julian Alps, NW Slovenia). Sediment. Geol. 165, 265­283. Bonnefoy-Claudet, S., Cornou, C., Bard, P.Y., Cotton, F., Moczo, P., Kristek, J., Fäh, D., 2006. H/V ratio: a tool for site effects evaluation. Results from 1-D noise simulations. Geophys. J. Int. 167, 827­837. Buser, S. 1986: Basic geological map of Yugoslavia 1:100.000 -- sheet Tolmin and Videm. Geological Survey of Slovenia. Fitzko, F., Suhadolc, P., Aoudia, A., Panza, G.F., 2005. Constraints on the location and mechanism of the 1511 Western Slovenia earthquake from active tectonics and modeling of macroseismic data. Tectonophysics 404, 77­90. Gallipoli, M.R., Mucciarelli, M., Castro, R.R., Mochavesi, G., Contri, P., 2004a. Structure, soil-structure response and effects of damage based on observations of horizontal-to-vertical spectral ratios of microtremors. Soil Dyn. Earthqu. Eng. 24, 487­495. Gallipoli, M.R., Mucciarelli, M., Eeri, M., Gallicchio, S., Tropeano, M., Lizza, C., 2004b. Horizontal to Vertical Spectral Ratio (HVSR) measurements in the area damaged by the 2002 Molise, Italy, earthquake. Earthq. Spectra 20/1, 81­93. Gosar, A., Stopar, R., Car, M., Mucciarelli, M., 2001. The earthquake on 12 April, 1998 in Krn mountains (Slovenia): ground motion amplification study using microtremors and modelling based on geophysical data. J. Appl. Geophys. 47/2, 153­167. Jurkovsek, B. 1986: Basic geological map of Yugoslavia 1:100.000 -- sheet Beljak and Ponteba. Geological Survey of Slovenia. Kastelic, V., Zivci, M., Pahor, J., Gosar, A., 2006. Seismotectonic characteristics of the 2004 earthquake in Krn mountains. Potresi v letu 2004, EARS, pp. 78­87. Kuscer, D., Grad, K., Nosan, A., Ogorelec, B., 1974. Geology of the Soca valley between Bovec and Kobarid. Geologija 17, 425­476 (in Slovenian, with English summary). Lapajne, J., Sket Motnikar, B., Zupancic, P., 2001. Design ground acceleration map of Slovenia. Potresi v letu 1999, pp. 40­49. EARS (in Slovenian, with English summary).

A. Gosar / Engineering Geology 91 (2007) 178­193 Mucciarelli, M., Gallipoli, M.R., 2001. A critical review of 10 years of microtremor HVSR technique. Boll. Geofis. Teor. Appl. 42, 255­266. Mucciarelli, M., Monachesi, G., 1998. A quick survey of local amplifications and their correlation with damage observed during the Umbro­Marchesan (Italy) earthquake of September 26, 1997. J. Earthqu. Eng. 2/2, 325­337. Mucciarelli, M., Contri, P., Mochavesi, G., Calvano, G., Gallipoli, M.R., 2001. An empirical method to assess the seismic vulnerability of existing buildings using the HVSR technique. Pure Appl. Geophys. 158, 2635­2647. Mucciarelli, M., Gallipoli, M.R., Di Giacomo, D., Di Nota, F., Nino, E., 2005. The influence of wind on measurements of seismic noise. Geophys. J. Int. 161, 303­308. Nakamura, Y., 2000. Clear identification of fundamental idea of Nakamura's technique and its applications. 8 pp., Proceeding of the 12th WCEE, Auckland, New Zealand. Natale, M., Nunziata, C., 2004. Spectral amplification effects at Selllano, Central Italy, for the 1997­98 Umbria seismic sequence. Nat. Hazards 33, 365­378. Panou, A.A., Theodulidis, N., Hatzidimitriou, P., Stylianidis, K., Papazachos, C.B., 2004. Ambient noise horizontal-to-vertical spectral ratio for assessing site effects estimation and correlation with seismic damage distribution in urban environment: the case of city of Thessaloniki (Northern Greece). Soil Dyn. Earthqu. Eng. 261­274. Parolai, S., Galiana-Merino, J.J., 2006. Effects of transient seismic noise on estimates of H/V spectral ratios. Bull. Seismol. Soc. Am. 96/1, 228­236.

193

Perniola, B., Bressan, G., Pondrelli, S., 2004. Changes in failure stress and stress transfer during the 1976­77 Friuli earthquake sequence. Geophys. J. Int. 156, 297­306. Ribicic, M., Vidrih, R., Godec, M., 2000. Seismogeological and geotechnical conditions of buildings in Upper Soca territory, Slovenia. Geologija 43, 115­143 (in Slovenian, with English summary). SESAME, 2004. Guidelines for the implementation of the H/V spectral ratio technique on ambient vibrations: measurements, processing and interpretation. http://sesame-fp5.obs.ujf%1Egrenoble. fr/Delivrables/Del%1ED23%1EHV_User_Guidelines.pdf, 62 pp. Vanini, M., Villani, M., Faccioli, E., Gosar, A., 2006. Modelling of strong ground motion of the July 2004, Mw 5.2 earthquake in Krn mountains. 3rd International Symposium on the Effects of Surface Geology on Seismic Motion, vol. 2. 10 pp., LCPC, Grenoble. Yuen, K.V., Beck, J.L., Katafygiotis, L.S., 2002. Probabilistic approach for modal identification using non-stationary noisy response measurements only. Earthq. Eng. Struct. Dyn. 31/4, 1007­1023. Zivci, M., Krn-2004 team, 2006. The Krn mountains (Slovenia) MW 5.2 earthquake: data acquisition and preliminary results. Geophys. Res. Abstr. 8, 06439. Zupancic, P., Ceci, I., Gosar, A., Placer, L., Poljak, M., Zivci, M., 2001. The earthquake of 12 April 1998 in the Krn Mountains (Upper Soca valley, Slovenia) and its seismotectonic characteristics. Geologija 44, 169­192.

Information

doi:10.1016/j.enggeo.2007.01.008

16 pages

Find more like this

Report File (DMCA)

Our content is added by our users. We aim to remove reported files within 1 working day. Please use this link to notify us:

Report this file as copyright or inappropriate

393504


You might also be interested in

BETA
doi:10.1016/j.enggeo.2007.01.008